Nutritionix_mcp

What is Nutritionix_mcp?

Nutritionix_mcp is a prototype chat application that lets users log meals in natural language. It parses food items via an MCP server backed by the Nutritionix API and returns detailed calorie and macronutrient breakdowns. It uses FastAPI for backend with WebSocket and chat endpoints, FastMCP to expose Nutritionix tools, LangChain-Ollama for tool‐calling and reasoning, and Google Mesop for the frontend.

How to use Nutritionix_mcp?

Users interact with the chat UI by typing natural language meal descriptions (e.g., “I had 1 katori dahi and 2 chapatis”). The system automatically calls Nutritionix tools to parse the food items and returns a breakdown of calories, protein, carbs, and fats. The LLM orchestrator can be swapped by modifying llm_provider.py to use different Ollama models.

Key features of Nutritionix_mcp

• Natural-language meal logging
• Automatic tool-calling to Nutritionix natural nutrients API
• Regional unit conversion via alt_measures
• Macro tracking: calories, protein, carbs, fats
• Modular LLM orchestration (swappable Ollama models)
• Lightweight Python UI using Mesop’s mel.chat

Use cases of Nutritionix_mcp

• Logging meals by describing them in everyday language
• Getting instant calorie and macronutrient breakdowns for food items
• Converting regional units (e.g., ounces to litres) automatically
• Experimenting with different LLM models for meal parsing and reasoning

FAQ from Nutritionix_mcp

What data sources does Nutritionix_mcp use?

It uses the Nutritionix natural nutrients API to parse food items and retrieve nutritional information.

What runtime dependencies does Nutritionix_mcp require?

It requires FastAPI, FastMCP, LangChain-Ollama, Google Mesop, and the Nutritionix API credentials.

How does the chat communicate with the backend?

The backend exposes WebSocket and chat endpoints (FastAPI) that the Mesop frontend connects to.

Can I use a different LLM model?

Yes, the LLM orchestrator is modular; you can swap Ollama models by editing llm_provider.py.

Does Nutritionix_mcp support regional units?

Yes, it uses alt_measures from the Nutritionix API to handle units like ounces and litres.